Y'-Help1, a DNA helicase encoded by the yeast subtelomeric Y' element, is induced in survivors defective for telomerase.

The yeast Y' element is a highly polymorphic repetitive sequence present in the subtelomeric regions of many yeast telomeres. The Y' element is classed as either Y'-L or Y'-S, depending on its length. It has been reported that survivors arising from telomerase-deficient yeast mutants compensate for telomere loss by the amplification of Y' elements. The total Saccharomyces cerevisiae genome DNA data base was searched for Y' elements, and 11 Y'-Ls and eight Y'-Ss were identified. As reported previously, many of the sequences were found to contain long open reading frames which potentially encode helicase. We examined the expression of the Y' elements in telomerase-deficient Deltatlc1 survivors, in which the TLC1 gene encoding the yeast telomerase template RNA had been disrupted, and found that the Y' element is highly expressed in the survivors, but not in the wild-type cells. Moreover, we demonstrated that the survivors produce a Y'-encoded protein designated as Y'-Help1 (Y'-helicase protein 1), and that this protein possesses helicase activity. Therefore, we suggest that the Y' element has a novel and potentially important role in trans, in addition to the well characterized role in cis, in telomerase-independent telomere maintenance in yeast.

During conventional DNA synthesis, the length of the chromosomal ends or telomeres is shortened with each cell division due to the end replication problem (reviewed in Ref. 1). Telomerase, a specialized form of reverse transcriptase that elongates the telomeric DNA, has been shown to play a primary role in maintaining functional telomeres (reviewed in Ref. 2). It is also known that telomerase-independent telomere maintenance mechanisms operate under certain circumstances (reviewed in Ref. 3). Indeed, Drosophila melanogaster completely lacks the canonical TG-rich telomeric repeats that are synthesized by telomerase. In this insect, retroposons called HeT-A and TART transpose to and elongate the chromosomal ends (4). In humans, a small but significant number of established human cell lines and clinical cancers have been found not to contain any detectable telomerase activity, although they maintain their telomeric TTAGGG repeats during extended proliferation (5,6). Therefore, the presence of telomerase-independent mecha-nisms of telomere elongation, which are referred to as ALT (alternative lengthening of telomeres) (7), was suggested. Retrotransposition similar to that found in Drosophila is considered an unlikely mechanism for human alternative lengthening of telomeres, because authentic human TTAGGG repeats have been observed, and no retroposon-like sequence has been found in the human alternative lengthening of telomeres (reviewed in Ref. 8). Thus, divergent molecular mechanisms appear responsible for telomerase-independent telomere maintenance in different species.
In Saccharomyces cerevisiae, the terminus of the telomeric DNA consists of telomeric repeats of (TG 1-3 ) n (9). The size of these clusters of repeats varies among different chromosomes and strains, but it is maintained within a relatively narrow range, the average being typically about 300 nucleotides. Several yeast genes are involved in telomere maintenance (10 -13). Of these, mutations in the EST1 gene result in a progressive loss of telomere sequence. The est1 mutants stop growing approximately 50 generations after EST1 disruption due to complete loss of the telomeres. This characteristic delayed phenotype has been referred to as yeast "senescence" (14). However, during further culture of the est1 mutants, a subpopulation that had resumed growth spontaneously appeared at a low but significant rate (15). These survivors still had very short stretches of terminal telomeric repeats, but, interestingly, showed high levels of amplification of the YЈ element, a repetitive sequence found in many yeast subtelomeric regions.
Two subtelomeric repeats called YЈ and X are frequently found associated with the telomeric repeats (16). YЈ is located adjacent to the telomeric repeats, either as a single copy or as a tandem repeat of two to four copies, at many but not all telomeres (Fig. 1). Two types of YЈ elements are known, YЈ-S elements are 5.2 kb 1 and YЈ-L elements are 6.7 kb in length, respectively (16 -18). Any particular tandem array of YЈ elements consists of either YЈ-S or YЈ-L, but not a combination of both elements (17). The sizes and sequences of the YЈ-Ss and YЈ-Ls are well conserved among the different strains, but their copy numbers and locations are highly polymorphic (16,17,19). Typically, 10 to 30 YЈ elements are present at approximately half the number of telomeres in a haploid genome (16,17). Another group of subtelomeric repeats called the X elements, which are much less conserved in size and sequence, are found at regions centromeric to YЈ elements. Short stretches of (TG 1-3) n -repeats are present at the boundaries of successive tandem YЈ elements, and at some boundaries of X and Y' elements (20,21). Both YЈ and X elements appear non-essential for telomere functions in cis, because some telomeres do not contain these elements (22,23).
Genetic studies have indicated that the EST1 gene belongs to the same epistasis group as the EST2, 3, and EST4/CDC13 genes and the TLC1 gene (24). TLC1 and EST2 encode the template RNA component and the catalytic subunit of yeast telomerase, respectively (13,25). Deletion of the TLC1 or EST2 gene results in the loss of telomerase activity (13,26), cell growth arrest at about 75 generations after the deletion (senescence), and the appearance of survivors that maintain telomeres by telomerase-independent mechanisms (13,25). The strikingly similar phenotypes observed with the est1, est2, and tlc1 mutants, along with genetic studies, suggest that these genes are involved in the same pathway, namely telomere maintenance by telomerase. Thus, a high degree of YЈ element amplification characterizes survivors that have arisen from yeast mutants defective for the telomerase pathway, presumably as a means of compensating for the loss of telomeric repeats and to stabilize chromosomes in telomerase-defective survivors. These amplified YЈ elements are unstable: the magnitudes of amplification and growth rates fluctuates during cell passages, even when they were derived from a single colony. The appearance of these survivors depends on RAD52, a gene essential for homologous recombination in yeast (15). These results suggest that the YЈ element was amplified by homologous DNA recombination. Interestingly, the YЈ elements encode a long open reading frame (ORF), potentially encoding amino acid sequences with putative helicase motifs (18). However, the significance of this open reading frame in the telomerase-independent telomere maintenance mechanisms is not understood. Therefore, we undertook a biochemical study focusing on the potential roles of the YЈ ORF in the establishment of survivors from the tlc1 deletion mutant.

EXPERIMENTAL PROCEDURES
Gene Disruption of the TLC1 Gene-The plasmid pYT211 harboring the TLC1 gene was a gift from Dr. Y. Ohya. The 4.4-kb BamHI-SphI fragment, which contains TLC1 and the neighboring gene CLS2, was cloned into pBluescript SK(ϩ) (Stratagene) to generate pBS/TLC1. The BamHI-PstI LEU2 fragment from pJJ282 (27) was inserted in place of the pBS/TLC1 BglII-NsiI fragment to generate ⌬tlc1::LEU2, in which most of TLC1 (89%) was replaced by the LEU2 marker gene, leaving 146 nt at the 3Ј terminus. TLC1 disruptants were constructed by transforming the diploid strain YPH501 (28) with the purified BamHI-HindIII fragment of ⌬tlc1::LEU2. Disruption of the TLC1 gene was confirmed by Southern hybridization with DNA derived from the transformed diploids and the resulting haploid segregants. One diploid strain, called YMY1, was found to have disruptions at both TLC1 alleles. Thus, YMY1 was a diploid strain homozygotic for ⌬tlc1::LEU2/⌬tlc1::LEU2 with no wild-type TLC1 gene. Another diploid strain called YMY2 was TLC1/⌬tlc1::LEU2 with one wild-type gene and one disrupted TLC1 gene.
Southern Analysis-Yeast total genomic DNA was isolated according to Kaiser et al. (29). Genomic DNA was digested with XhoI for the analysis of telomere length and YЈ element amplification. Yeast telomeres were detected using 32 P-end-labeled poly(dGT) 15 as a probe. The YЈ element and the URA3 gene were detected using the XbaI-HindIII fragment of the YCF4 plasmid (30), and the SmaI-PstI fragment of the pJJ242 plasmid (27) as probes, respectively. Hybridization with radioactive probes was carried out according to the manufacturer's instructions (Amersham Pharmacia Biotech). The blots were exposed to x-ray film, or to a Fuji imaging plate, and analyzed using a Bioimaging Analyzer BAS2000 (Fuji Film).
RNA Isolation-Total RNA was prepared by a modification of the procedure of Schmitt et al. (31). Poly(A) ϩ mRNA was isolated from total RNA by oligo(dT)-cellulose affinity chromatography as described (32).
Northern Analysis-Total RNA or poly(A) ϩ RNA was separated by electrophoresis in a 1% agarose gel containing 2.2 M formaldehyde and transferred to a GeneScreen membrane (DuPont). ACT1, TLC1, and the transcript product of the YЈ region were detected using probes using the XhoI-HindIII fragment of pYA301 (33), the BamHI-HindIII fragment of pBS/TLC1, and the XbaI-HindIII fragment of YCF4, respectively (30).
Cloning of YЈ-Help1-A YMY1 (⌬tlc1/⌬tlc1) cDNA library was constructed from 6 g of poly(A) ϩ RNA using the ZAP-cDNA synthesis kit (Stratagene). The XbaI-HindIII fragment of YCF4 was used to screen 6 ϫ 10 4 independent clones from the library. Eighteen positive clones (YЈc01-c18) were isolated. Further screening of another 4 ϫ 10 4 independent clones using the YЈc17 cDNA, which has the longest 5Ј extension of the cDNAs isolated, gave rise to seven positive clones (YЈ2c01-c07). Three of 25 clones (YЈc12, YЈ2c03, and YЈ2c05) had the same 5Ј-end sequence and contained a poly(A) ϩ tract, which suggested that these clones contain the full-length cDNA.
Antibody Production-The recombinant YЈ-Help1 protein was prepared in Escherichia coli transformed by pGEX (Amersham Pharmacia Biotech) or pET30 (Novagen) expression vectors expressing a fusion protein corresponding to amino acid 50 -234 (HelpN) of the YЈc12 ORF. Three Japanese White rabbits (female, 2 kg weight) were subcutaneously injected with 100 g of the affinity purified His-tagged HelpN fusion protein mixed with Freund's complete adjuvant. The anti-HelpN antibodies were affinity purified through a column conjugated with the GST-HelpN fusion protein, and were used for Western analysis. The YЈc12 cDNA, which contains the entire YЈ-Help1 ORF, was inserted into the EcoRI site of pYES2, a yeast expression vector (Invitrogen). The resulting pYES2/YЈ plasmid was introduced into yeast and the recombinant YЈ-Help1 protein was used as a positive control in Western blotting experiments.
Western Analysis-Whole cell extracts were prepared according to Lue et al. (34). Whole cell lysates from mutant or wild-type cells were separated on an 8% SDS-PAGE, and transferred to a polyvinylidene FIG. 1. Telomere and subtelomere structures in S. cerevisiae. Two types of YЈ elements with lengths of 5.2 and 6.7 kb (shown as boxes) are shown as tandem repeats bordering the terminal telomeric TG-repeats (shown as the terminal arrowheads). Internal TG repeats are present between two successive YЈ elements and between YЈ and the internal regions (shown as the internal arrowheads). XhoI digests DNA at a single site in the YЈ elements, but not within the telomeric repeats. Southern hybridization of XhoI restricted yeast genomic DNA using a sequence of TG repeats as a probe detects: 1) fragments of terminal telomeric repeats associated with a part of the YЈ element (this fragment is typically 1.2 kb); 2) fragments corresponding to a single unit of YЈ element derived from two successive YЈ repeats (5.2 and 6.7 kb); and 3) fragments derived from the most internal YЈ element and adjacent unique sequences (heterogeneous sizes). difluoride membrane (Millipore). Western blotting was carried out as described in Ref. 35, using 1 g/ml anti-HelpN antibody and a 1/1000 dilution of horseradish peroxidase-linked donkey anti-rabbit immunoglobulin (Amersham Pharmacia Biotech). Specific protein bound to the membranes was detected using the ECL system (Amersham Pharmacia Biotech).
Baculovirus Transfer Vectors-A BamHI-SalI fragment of YЈc12 containing the entire YЈ-Help1 sequence was inserted into the BamHI-SalI cloning sites of the pFASTBAC HTa expression vector (Life Technologies). The resulting pFASTBAC HT/Help was introduced into DH10BAC E. coli cells (Life Technologies), and the insert was transposed onto the bacmid present in the DH10BAC cells. The recombinant bacmid was transfected into Sf9 cells. The recombinant virus was isolated after protein analysis by Western blotting using the YЈ-Helpspecific antibody.
Extraction of the Soluble Protein Fraction of Sf9 Cells-The Sf9 cells were washed once with cold phosphate-buffered saline 3 days after infection. The infected cells were pelleted by centrifugation (500 ϫ g for 5 min at 4°C) and stored at Ϫ80°C until use. To purify the protein, the frozen cells were thawed and suspended in 5 volumes of lysis buffer containing 50 mM Tris-HCl (pH 8.5), 10 mM 2-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, and 1% Nonidet P-40. After rotating the tube for 1 min, the mixture was centrifuged at 10,000 ϫ g for 10 min to remove cell debris. The cleared supernatant was used in the biochemical experiments.
DNA Helicase Assay-DNA helicase activity was measured using the release of the 5Ј-32 P-labeled 17-mer DNA fragment 5Ј-GTAAAAC-GACGGCCAGT-3Ј from an annealed circular M13mp18 DNA molecule (36). The labeled DNA substrate (1000 cpm) was incubated at 37°C for 20 min with the soluble protein fraction (about 5 g) in a reaction mixture (10 l) containing 20 mM Tris-HCl (pH 7.5), 8 mM dithiothreitol, 2 mM MgCl 2 , 2 mM ATP, 10 mM KCl, 4% (w/v) sucrose, and 80 g/ml bovine serum albumin. The reaction was terminated by addition of a final concentration of 0.006% SDS, 2 mM EDTA, 1% glycerol, and 0.02 mg/ml bromphenol blue. After further incubation at 37°C for 5 min, the substrate and product were resolved by electrophoresis through a 12% polyacrylamide gel and visualized by autoradiography of the dried gel.
Data Base Search for the YЈ Element in the Yeast Total Genomic Sequence-The YЈ-L sequence completely sequenced by Louis and Haber (Ref. 18; GenBank accession number U23472; located at the chromosome XV right telomere in YP1 yeast strain) was used as a YЈ-L probe to retrieve related sequences from the data base. To obtain YЈ-S, GenBank was first searched using the incomplete YЈ-S sequence reported by Louis and Haber (18). This procedure identified a complete YЈ-S sequence (EMBL Y08934), and this sequence was then used as a data base search probe. These sequences were then used to search the yeast complete genome nucleotide sequence in the Stanford Saccharomyces Genome Data base (SGD) for YЈ-L and YЈ-S sequences using the BLAST program.

Data Base Search for YЈ-S and YЈ-L Elements in the Complete
Yeast Genome Sequence-We are interested in the possible roles of YЈ elements in producing survivors among telomerasedefective mutants. In a previous study by Louis and Haber (18), the nucleotide sequences and polymorphisms of YЈ-L and YЈ-S were examined in detail. However, they did not report the complete nucleotide sequences of YЈ-S. Since the publication of their study, the complete nucleotide sequence of the yeast genome has become available. Therefore, we systematically searched the Saccharomyces Genome Data base (SGD) for yeast YЈ-S and YЈ-L elements, and identified a total of 11 YЈ-Ls and eight YЈ-Ss in the yeast strain S288C that was used for genome sequencing. However, this number of YЈ elements should be considered as the minimum, since the current yeast genome data base does not contain all telomeric regions. The detailed results of this analysis will be reported elsewhere.
YЈ-Ls and YЈ-Ss share common nucleotide sequences along their lengths. YЈ-Ss contain several deletions compared with YЈ-Ls, most of which are found in the 5Ј-regions of the sequences. These deletions account for the difference in length between YЈ-S and YЈ-L (about 5.2 versus 6.7 kb). A number of polymorphisms, including base substitutions, small insertions and deletions, were found among different members of the YЈ-S and YЈ-L groups. A schematic diagram showing the YЈ-L and YЈ-S structures is shown in Fig. 2. As reported in the previous study (18), several long ORFs were present in all YЈ elements. The numbers and lengths of these ORFs varied among the different members due to polymorphism. However, the following features are shared by many members (Fig. 2). In the YЈ-L group, a single long ORF encoding about 1796 amino acids was present in six out of 11 YЈ-L members. Although the previous study identified two overlapping ORFs instead of one long continuous ORF, only two of our 11 clones showed this type of ORF composition. In the YЈ-S group, one long open reading frame encoding about 997 amino acids was present in most members. Since the sequences of YЈ-S and YЈ-L are almost identical in the 3Ј-region, the amino acid sequence encoded by the YЈ-S ORF is included in the putative YЈ-L amino acid sequence in its C-terminal region. As reported in the previous study, a group of motifs characteristic of many DNA or RNA helicases was found in these common regions (Fig. 2).

YЈ Element Amplification in ⌬tlc1
Survivors-The diploid strain YPH501 was transformed with ⌬tlc1::LEU2, producing YMY2 cells heterozygous for ⌬tlc1 (see "Experimental Procedures"). After meiosis, four viable spores were produced from YMY2, indicating that spores bearing a TLC1 gene disruption are viable, as previously reported (25). They were cultured in rich media and examined for telomere length at intervals up to about 120 generations after haploid cell production. In this experiment, the isolated genomic DNA was digested with XhoI, which cuts the subtelomeric YЈ element at a single site, leaving the telomeric repeats undigested. The DNA was then analyzed by Southern hybridization using TG probes to detect telomeric repetitive sequences (see Fig. 1). In the wild type haploid YPH500 cells, an intense signal of about 1.2 kb was observed that represents the telomere restriction fragments (TRFs) derived from YЈ-containing telomeres (Fig. 3A, lane 1). Telomeric DNAs derived from non-YЈ-containing telomeres were represented as several TRFs migrating more slowly than the 1.2-kb band. The TRFs in the TLC1 haploid cells did not shorten during culture up to 120 generations (data not shown). However, the ⌬tlc1 haploid cells showed a gradual reduction in the size and intensity of TRFs, as indicated in Fig. 3A, lanes 2-8. The ⌬tlc1 cells grew as well as the wild-type cells during the early passages. However, these cells gradually grew more slowly, and the plating efficiencies declined. In a typical case, after 50 generations, most of the plated cells formed very tiny or no colonies. These experiments confirmed the results obtained by others showing that ⌬tlc1 cells show a senescence phenotype (25).
During prolonged cultures, cells that apparently grew stably occasionally appeared. These growth-resuming cells most likely corresponded to the survivor cells that had been originally described for the est1 mutant (14,15). Four independent ⌬tlc1 survivor strains were analyzed for telomere restriction fragments (Fig. 3A, lanes 9 -12). These cells had essentially lost all TRFs, but characteristically possessed highly amplified signals at 5.2 and 6.7 kb, which were also found in the wild-type and the ⌬tlc1 cells during the earlier passages, albeit with reduced intensities. These bands also hybridized with a YЈspecific probe (Fig. 3B). Since the YЈ elements are bordered by telomeric TG 1-3 -repeats, and are digested by XhoI at a single site, Southern blot analysis using either the TG-repeat probe or the YЈ probe detects YЈ-derived bands corresponding to a single unit size of either 5.2 or 6.7 kb (Fig. 1). Therefore, it was concluded that the YЈ element is amplified in the ⌬tlc1 survivors, as reported for the est1 survivors (15). Taken together, these results indicated that the TLC1 deletion resulted in telomere loss, a senescence phenotype, and the spontaneous appearance of survivors demonstrating YЈ amplification. This result suggested that this set of phenotypes might be shared by a null mutant for any gene of the EST1-epistasis group. As shown in previous studies (15), survivors did not appear in ⌬tlc1 rad52 mutants, suggesting that YЈ amplification occurred by homologous recombination.
Expression of YЈ Elements in ⌬tlc1 Survivors-The data base analysis indicated that most YЈ elements contain long ORFs. Significantly, the amino acid sequences predicted from these ORFs contain a helicase-like motif. Therefore, it was of interest to determine whether the amplified YЈ elements in the ⌬tlc1 survivors were expressed. Total RNAs were isolated from the TLC1 and ⌬tlc1 survivors and subjected to Northern analysis to examine expression of the TLC1, YЈ element, and ACT1 genes. As shown in Fig. 4A, the ACT1 gene mRNA was detected at similar levels in all RNA samples, which ensured that the amounts of RNA loaded in the Northern analysis were similar (lanes 1 and 2). The TLC1 gene was expressed in TLC1 cells but not in the ⌬tlc1 survivors, as expected (lanes 3 and 4). In an earlier study, a low level of YЈ element expression was detected in wild-type cells by a RNase protection assay but not by Northern blot analysis (18). We also did not detect any YЈ transcript in TLC1 cells by Northern blot analysis (lane 5). However, the ⌬tlc1 haploid survivors showed a significant level of YЈ expression that could be easily detected by this method (lane 6). We concluded that YЈ element transcription is upregulated in the ⌬tlc1 survivors. When poly(A) ϩ RNAs were examined, three YЈ transcripts of 6.8, 5.1, and 3.4 kb were clearly distinguishable in the ⌬tlc1 haploid survivors, with the 3.4-kb transcript being most abundant (Fig. 4B).
Since we did not know a priori which alleles among the many copies present in the yeast strain we had studied were amplified and expressed in the ⌬tlc1 cells, we cloned the YЈ element cDNA. A cDNA library was constructed using poly(A) ϩ RNA derived from ⌬tlc1/⌬tlc1 diploid survivors that had also shown a high degree of YЈ element amplification and expression (data not shown). Twenty-five clones were identified using a YЈspecific probe from 1 ϫ 10 5 independent cDNA clones. A cDNA clone (YЈc12) representing the full-length transcript starting around the single EcoRI site (see Fig. 2) was completely sequenced. The sequence contained a poly(A) ϩ tract at its 3Ј terminus, indicating that YЈ elements are transcribed as a poly(A) ϩ RNA. A potential poly(A) addition signal, AAGAAA, was found 16 nt upstream from the poly(A) ϩ sequences. YЈc12 contained a nucleotide sequence of 3104 base pairs ϩ poly(A), a value close to the 3.4-kb length characteristic of the most abun- FIG. 3. Telomeres and YЈ elements in the wild-type and the ⌬tlc1 survivors cells. A, XhoI-digested genomic yeast DNAs (10 g each) were hybridized with the d(TG) 15 probe. Lane 1, the wild-type YPH500 cells; lanes 2-8, a ⌬tlc1 strain grown for 10, 30, 50, 70, 80, 100, and 120 generations after tetrad dissection; lanes 9 -12, four independent ⌬tlc1 survivors isolated from the ⌬tlc1 strain. The TRF was gradually shortened as the ⌬tlc1 cells grew. YЈ-TRF indicates the TRFs derived from the YЈ-containing telomeres. YЈ-L (6.7 kb) and YЈ-S (5.2 kb) were highly amplified in the survivors (arrowheads on the right). B, XhoI-digested genomic yeast DNAs (10 g each) derived from the wild-type YPH500 (lanes 1 and 3) and ⌬tlc1 survivor cells (lanes 2 and 4) were hybridized with the d(TG) 15 probe (lanes 1 and 2), or the YЈ-specific probe (lanes 3 and 4). The TRF positions are shown by open arrowheads. The YЈ-L and YЈ-S were hybridized both with the d(TG) 15 probe and the YЈ-specific probe, as indicated by filled arrowheads. dant YЈ transcript deduced by the Northern analysis, suggesting that the major cDNA clones might correspond to the fulllength 3.4-kb transcript. The cDNA sequence was almost identical to the genomic YЈ-L and YЈ-S sequences identified from the data base search. Therefore, the YЈ transcript apparently does not undergo splicing. At present, we do not know whether this transcript was derived from YЈ-S, YЈ-L, or both. Interestingly, the coding region downstream from the EcoRI site was found to be most conserved in YЈ elements identified from the complete yeast genome sequence (data not shown). It was likely that one cDNA clone (YЈc15), in which the 5Ј terminus extended beyond the majority of cDNAs, was derived from the longer 6.8-and 5.1-kb transcripts. The relatively lower abundance of this type of cDNA might reflect the lower abundance of longer transcripts.
The YЈc12 clone contained a long ORF. This ORF continued to the 5Ј terminus without termination by a stop codon. Assuming that the first ATG codon is used as the initiation codon, a polypeptide consisting of 997 amino acids was predicted, as shown in Fig. 5. The predicted molecular mass of the product is 112 kDa. As will be described later, this putative YЈ protein was actually translated in survivors and had helicase activity. Therefore, we will refer to this protein product as YЈ-Help1 (standing for YЈ-helicase protein 1). The N-terminal region of the YЈ-Help1 sequence contained the seven motifs (motif I, Ia through VI) conserved in the DEAD/DEAH or DEXH subfamily of helicases (37-41), which have been described previously (18). The helicase motif region of YЈ-Help1 was analyzed by the BLAST program (42) for GenBank. The RNA helicase, eukaryotic translation initiation factor 4A (43), scored the highest homology value (E value ϭ 3 ϫ 10 Ϫ5 -1 ϫ 10 Ϫ4 ). However, the SAT motif found in motif III of many RNA helicases (44,45) was present in eukaryotic initiation factor-4A but not in YЈ-Help1. Another gene of interest that showed a high homology value with YЈ-Help1 was the RecG protein of the hyperthermophilic bacterium, Aquifex (E value ϭ 4 ϫ 10 Ϫ4 ). RecG is a DNA helicase involved in homologous DNA recombination (46). Eleven tandem repeats of the serine-threonine-rich 12-amino acid sequence (SXTXATTTXSXX; underlined in Fig. 5) were found in the middle region of the YЈ-Help1 sequence. The number of 36-base pair tandem repeats encoding this sequence is known to vary among different YЈ genomic clones (18). This feature was also observed with our collection of YЈ elements from the complete yeast genome sequence (data not shown). The repeat number also varied among different cDNA clones identified in this study. This result indicated that the YЈ element was transcribed from more than one YЈ allele. A BLAST search did not identify any sequence with significant homology to this repetitive sequence in the GenBank data base.
The YЈ-Help1 Protein Is Present in the ⌬tlc1 Strain-We next examined whether the ORF present in the YЈ transcript was translated to produce the predicted peptide. A recombinant protein corresponding to motifs Ia through III of YЈ-Help1 was prepared in E. coli, and was used as an antigen to obtain a polyclonal antiserum against the putative YЈ-Help1 polypeptide. This antibody, designated as anti-HelpN, was affinitypurified and was tested for its reactivity with the recombinant YЈ-Help1. The YЈ-Help1 cDNA was inserted into the high copy number yeast expression plasmid pYES2 to produce pYES2/YЈ. This vector contains the GAL1 promoter to drive the expression of the insert. pYES2/YЈ was introduced into wild-type YPH500 cells, and the cells were cultured in the presence of galactose to express the recombinant protein. In a Western blotting experiment, the anti-HelpN antibody did not recognize any band in the wild-type haploid cell YPH500 extract (Fig. 6, lane 1), but detected a prominent band in the pYES2/YЈ-transformed YPH500 cells (lane 2). The apparent molecular mass of this protein calculated from its migration in SDS-PAGE (132 kDa) was larger than the mass predicted from the amino acid sequence of YЈ-Help1 (112 kDa). These results indicated that the anti-HelpN antibody specifically recognized the YЈ-Help1 protein. We next examined cell extracts obtained from the ⌬tlc1 survivor cells, in which the YЈ element had been amplified after extensive cultivation. No YЈ-Help1 was detected by this assay using the TLC1 haploid cells (lane 3). In contrast, in the ⌬tlc1 survivor cells, significant amounts of YЈ-Help1 that showed the same mobility as the recombinant YЈ-Help1 produced by pYES2/YЈ (lane 2) were detected (lane 4). In conclusion, YЈ-Help1 was not present in the wild-type yeast cells, but was produced in ⌬tlc1 survivors.
The YЈ-Help1 Protein Catalyzes DNA Unwinding-To analyze the biochemical properties of the YЈ-Help1 protein, recombinant protein was prepared in Sf9 insect cells using a baculovirus expression system. The lysates of cells infected by either the virus without the YЈ element or the recombinant virus expressing the YЈ-Help1 cDNA were analyzed by Western blotting using the anti-HelpN antibody. Recombinant YЈ-Help1 protein was detected specifically in the cDNA-infected cells but not in vector-infected cells (Fig. 7A). The mobility of this baculovirus recombinant protein was similar to that of the yeast recombinant and the native YЈ-Help1. Helicase activity was examined using these Sf9 cell extracts. A 5Ј-end-labeled oligonucleotide annealed to single-stranded M13 DNA was used as a substrate to detect the putative helicase activity of the YЈ-Help1 expressed by the Sf9 cells (Fig. 7B). Under our experimental conditions, no intrinsic helicase activity was detected with the vector-transfected Sf9 cells. In contrast, all three extracts obtained from independent YЈ-Help1-expressing Sf9 cells showed a significant level of helicase activity, as assessed by dissociation of the oligonucleotide from M13 DNA. This data indicates that YЈ-Help1 is a DNA helicase. DISCUSSION Progressive telomere shortening, senescence, and the appearance of survivors having a high degree of YЈ element amplification have been reported with est1, est2, and tlc1 mutants (15, 24, 25, this study). In these cases, YЈ amplification seems to occur in all telomeres presumably by DNA recombination, compensating for the telomere loss in the survivor cells. Therefore, the in cis role of YЈ element amplification in yeast telomeraseindependent telomere maintenance is well recognized. However, given that most YЈ elements contain long ORFs that potentially encode helicases, it is possible that YЈ element amplification also contributes in trans to telomere mainte- FIG. 4. Northern analysis of YЈ-derived transcripts. A, total RNAs (10 g each) obtained from the wild strain YPH500 (lanes 1, 3, and 5) and the ⌬tlc1 survivors (lanes 2, 4, and 6) were analyzed for the expression of the ACT1 gene (lanes 1 and 2), TLC1 (lanes 3 and 4) and the YЈ element ( lanes 5 and 6). B, 1 g of poly(A) ϩ RNAs obtained from the ⌬tlc1 survivors was analyzed for the expression of the YЈ element. nance in survivors, a possibility that has not been studied so far. In this study, we characterized the YЈ element biochemically, and showed for the first time that the amplified YЈ elements are actively transcribed in the ⌬tlc1 survivor cells, producing the YЈ-Help1 protein. Moreover, we demonstrated that YЈ-Help1 has helicase activity. These results suggest that YЈ-Help1 expression might have a novel in trans role in the survivor cells to compensate for telomere shortening by telomerase-independent mechanisms.
Since mutations in RAD52, a gene essential for homologous DNA recombination in yeast, result in a failure to produce survivor cells in telomerase-deficient cells, it has been proposed that YЈ element amplification is achieved by homologous DNA recombination (15). The idea that the yeast YЈ elements and the terminal telomeric repeats are amplified by DNA recombination is not novel, it was first proposed in 1984 (19,47). The YЈ element, which is known to contain an autonomously replicating sequence (48), was later found to exist episomally as au-tonomously replicating circular DNAs in addition to the genomic copies. Therefore, it is possible that YЈ element amplification can occur by a gene conversion mechanism among the many YЈ element copies in yeast, present either in the genome or episomally. The net increase in the copy number of the YЈ elements at telomeres presumably compensates for the progressive loss of telomeres in the absence of telomerase.
How could the YЈ-encoded helicase compensate for the telomere loss in telomerase-deficient cells? Several helicases are known to be involved in DNA homologous recombination (49 -  55). E. coli RuvABC proteins have been most extensively studied. The DNA helicase RuvB protein is responsible, with the aid of the RuvA protein, for the migration of Holliday junctions, extending heteroduplex regions during DNA recombination (49). Recently, it was reported that another E. coli DNA helicase RecQ, which is a homologue of human WRN and BLM helicases (50,54), also promotes homologous DNA recombination (52,55). Therefore, an attractive possibility is that YЈ-Help1 may facilitate YЈ amplification. In this study, we found that YЈ-Help1 is not expressed in telomere-proficient cells, but is highly induced in survivor cells that maintain telomeres in the absence of telomeric repeats. The expression of native YЈ elements in wild-type cells is repressed probably by the telomere silencing effect (56). However, once telomeres have shortened in telomerase-defective mutant cells and all telomeric repeats have been lost, the silencing effect is not effective. The YЈ element may then be derepressed to produce YЈ-Help1. YЈ-Help1 then may enhance homologous DNA recombination among YЈ elements and, as a consequence, may induce YЈ amplification to prevent chromosomal loss and cell death. The amplified YЈ elements will further increase the production of YЈ-Help1, establishing a positive feedback circuit. This mechanism will keep operating to counteract the telomere shortening unless YЈ-Help production is shut down by accidental loss of all YЈ elements at the telomeres. Since we did not see any YЈ element amplification in the YЈ-Help1 overproducing wild-type cells (data not shown), it seems that YЈ-Help1 can only be effective in certain circumstances, such as when the telomere and subtelomeric structures are altered due to telomere shortening. Also, it is important to realize that this back-up telomere maintenance mechanism is not physiological, since only a small fraction of cells among telomerase-deficient mutants can give rise to survivor cells. However, a previous study showed that the YЈ element is expressed during meiosis (57). It would be interesting to determine whether YЈ-Help1 helicase has a physiological role in telomere maintenance during meiosis.